EmbeddingGemma 2 and multimodal retrieval in one shared vector space
Radar: a post claims EmbeddingGemma 2, from Google, maps text, code, images, video and audio into a single vector space and uses an Apache 2.0 license. The source gives no metrics or technical documentation.
According to the Radar post dated 6 October 2026, EmbeddingGemma 2, from Google, maps text, code, images, video and audio into a single shared vector space. The same post states that the model uses the Apache 2.0 license. These are the post author's claims; the source presents no benchmarks, performance metrics or technical documentation.
The relevance lies in the possibility of simplifying retrieval across different media types, since queries and documents in different formats would become comparable in the same space. To verify, check the model's official documentation and the license text before any use. If your organization uses AI to study this material, do not paste internal documents, personal data or credentials into study tools; use public or anonymized excerpts.